http-client-procs.tcl

  • last updated 10 hours ago
Constraints
Constraints: committers
 
Constraints: files
Constraints: dates
Trim whitespace

Fix proc and contract doc elements, so they are properly parsed by apidoc::api_*_documentation.

- Add mising @param

- Add missing @author

- @creation_date -> @creation-date

- @creation-user -> @author

- Fix typos

  1. … 8 more files in changeset.
Avoid double encoding when content is in the same encoding as that of tcl interpreter

change capitalization for protocol name

improve spelling

  1. … 10 more files in changeset.
improve wording

  1. … 7 more files in changeset.
Revert massive replacement of empty list creation sentences. The use of '[list]' instead of '{}' adds semantics that could be used for performance improvements in the future, such as using a different internal representation. There is already work in this direction, avoiding the generation of the string representation during comparison of empty strings (huge thanks to Stefan Sobernig for the pointer: https://core.tcl.tk/tcl/info/44527c632ed609c2).

  1. … 475 more files in changeset.
Prefer '{}' to '[list]' when creating empty lists

  1. … 71 more files in changeset.
Fix typos

  1. … 42 more files in changeset.
Fix typos

  1. … 14 more files in changeset.
Fix even more typos

  1. … 29 more files in changeset.
Move deprecated procs into deprecated-procs.tcl

  1. … 10 more files in changeset.
Fix more typos

  1. … 11 more files in changeset.
Fix typos

  1. … 2 more files in changeset.
Bugfix

Decode form variables when a payload is transformed from form-urlencoded to multipart. Fix documentation for -formvars parameter

Bugfix

Avoid double encoding when content is in the same encoding as that of tcl interpreter

Fix nonsense fconfigure -translation $encoding

Try to make code clearer

merged changes from the oacs-5-9 branch and resolved conflicts

    • -404
    • +1012
    ./http-client-procs.tcl
  1. … 7834 more files in changeset.
Standardize spellings of names

  1. … 6 more files in changeset.
Standardize spelling of names of products (Tcl, AOLserver, PostgreSQL, NaviServer)

  1. … 43 more files in changeset.
Fix spelling errors

  1. … 22 more files in changeset.
Improve robustness of "file delete" operations

  1. … 25 more files in changeset.
Reinitialize file data array at each cycle

Fix typo, simplify and clarify content-type header manipulation

Don't add extra content-type when users specifies their own

Fix missing square bracket

Don't mess with user-defined content-type, unless required

Spool POST payload bigger than a threshold (default ~25MB) to a file, instead of keeping in memory the entire request.

This exploits (transparently) -body_file option in the ns_http API and native curl features.

Put some caching into utilities around to reduce db calls and execs.

Creation of this header is already done properly into ns_http

Insert content-length only when transfer-encoding is not chunked

- Aligned and fixed behavior for encoding of request with respect to redirect in the two implementations

- Use only binary encoding in curl implementation for the file used to spool the body, to avoid

corruption of content (eventual encoding has already happened in its proper place)

- Allow only for curl to exploit native file POSTing capability (only way to POST files > 2GB)

TODO: wrap this behavior in the generic API

- Add content-length header when POSTing content, as some servers will require it (curl did this already)